Download The Regulator of Body - Division Of Animal Sciences

The Regulator of Body
Temperature
I
-
-
m
••• :i
f
Brody Memorial Lecture VI
H. T. Hammel
Special Report 73
University of Missouri
I
August 3, 1966
Agricultural Experiment Station
THE BRODY LECTURESHIP COMMITTEE
Dr. William H. Pfander, Gamma Sigma Delta representative
Dr. Joseph E. Edmondson, Sigma XI representative.
Dr. Harold D. Johnson, Dept. ofDairy Husbandry, Chairman.
This Committee, appointed by Dean Elmer R. Kiehl, August,
1965, with joint sponsorship of the organizations these men repre
sent, plan at least one lecture a year.
DR. H. T. HAMMEL-BIOGRAPHY
H. T. Hammel was born in Indiana in 1921. He began his academic career at
Purdue University where he was graduated in physics in 1943. He was invited to the
Manhattan District's Los Alamos Laboratory where he worked in the Experimental
Physics Division on nuclear reactors.
While obtaining an M.S. degree in Physics at Cornell University, he was attracted
to Professor Donald R. Griffin from whom he subsequently obtained a Ph.D. in Zo
ology in 1953. Since graduating, his interests and development have ben strongly in
fluenced by two other distinguished physiologists, Professor P.P. Scholander, Scripps
Institution of Oceanography, and Professor James D. Hardy, University of Pennsyl
vania. He moved from Philadelphia to New Haven with Dr. Hardy and was ap
pointed Fellow of theJohn B. Pierce Foundation Laboratory and head of the work
ing group in physiology.
His major work has been in thermal physiology where his interests range from
thermal regulation in dogs, rats, monkeys, and hibernators to responses to thermal
stress in camel, reindeer, and jack rabbit; and on a cold adaptation in several ethnic
groups ofman including Australian Aborigines, Alacaluf Indians ofTierra del Fuego,
Kalahari Bushmen, Eskimos, and Norwegian youth.
His interests have been broadened by frequent collaborations with Dr. Scholander.
Most recently, with Dr. Scholander, he has developed a simple technique for measur
ing the negative hydrostatic pressure in the xylem sap of plants, providing direct evi
dence of tension in the sap of tall trees, mangrove trees, and desert plants.
Memberships: American Physical Society, American Society of Mammalogy;
American Physiological Society, American Society of Zoologists.
Some recent publications are:
Jackson, D. C., and H. T. Hammel. Hypothalamic "Set" Temperature Decreased in
Exercising Dog. Ufe Sciences No. 8, pp. 554-563, 1963.
Hammel, H. T., D. C. Jackson, J. A. J. Stolwijk.J. D. Hardy, and S. B. Stromme. Tem
perature Regulation by Hypothalamic Proportional Control with an Adjustable
Set Point. J. Appl. PhysioL, 18, No. 6:1146-54, Nov., 1963.
Hammel, H. T., S. B. Stromme and R. W. Cornew. Proportionality Constant for Hy
pothalamic Proportional Control ofMetabolism in Unanesthetized Dog. Life
Sciences, No. 12, pp. 933-947, 1963.
Chowers, I., H. T. Hammel, S. B. Stromme, and S. M. McCann. Comparison ofef
fect of environmental and preoptic cooling on plasma cortisol levels. The Ameri
can Journal ofPhysiology, Vol. 207, No. 3. Sept., 1964. pp. 577-582.
Scholander, P. F., H. T. Hammel, Edda D. Bradstreet, and E. A. Hemmingsen. Sap
Pressure in Vascular Plants. Science, April 16, 1965. Vol. l48, No. 3668, pp. 339346.
Hammel, H. T. Neurons and Temperature Regulation. Physiological Controls and
Regulations, Chapter 5. William S. Yamamoto and John R. Brobeck, Editors.
Hardy, J. D., J. A. J. Stolwijk, H. T. Hammel, and D. Murgatroyd. Skin Temperature
and cutaneous pain during warm water immersion. Joumel of Applied Physiology,
Vol. 20, No. 5, September 1965.
Abrams, R. M., J. A. J. Stolwijk, H. T. Hammel and H. Graichen. Brain Temperature
and Brain Blood Flow in Unanesthetized Rats. Life Sciences, Vol. 4, pp. 2399-2410
1965.
Schmidt-Nielsen, K., T. J. Dawson, H. T. Hammel, D. Hinds and D. C. Jackson.
The Jack Rabbit—a study in its desert survival. Hvalradets Skrifter, Nr. 48: pp.
125-142, 1965.
Hammel, H. T. Oxygen transport through hemoglobin solutions. Hvalradets Skrifter,
Nr. 48: pp. 164-177, 1965.
The Regulator of Body
Temperature
Presented February 17, 1966, University of Missouri
by
H. T. Hammel
John B. Pierce Foundation Laboratory
New Haven, Connecticut
There are many ways to begin a discussion of the regulation ofan animal's
body temperature, but on this occasion I think it is appropriate to begin with
a quotation from Bioenergetics and Growth (p. 265), "Homeothermy has many
aspects, theoretical, agricultural and engineering. The theoretical aspect is con
cerned with homeothermic mechanisms; the agricultural with the influence of
environmental temperature and humidity on productivities and efficiencies of
farm animals; the engineering with ventilation, heating and cooling." In a wry
note. Professor Brody continues—"We shall discuss each of these. The theoreti
cal and numerical discussions are presented in small type, the practical and gener
al in large type." My discussion tonight will be presented in very small type be
cause it is not only theoretical but also speculative.
I would like to orient you by reference to a generalized model of a regu
lating system as it pertains to body temperature, Figxire 1. You will quickly rec
ognize that Professor Brody and his colleagues have spent and his colleagues
are spending an enormous amount ofeffort working the right side ofthis model.
Oltturbing
Signol
Thermol
Stret*
CONTROLLING
CONTROLLED
SYSTEM
SYSTEM
—
Rtftrtnct
Rtlsranct
Input
Input
Adjusting
iiSOfl!
El*mtnt>
t'(il«ep)
Actuollng
±0 SIgnol
Regulated
Monlpulated
Variable
G(S)
Thermo-
•<tt|;
Thypo
reguldtory
Reiponte,
(R-R.)
next)
• tc.
Feedbock Signal
Fig. 1—Block diagram for regulation of hypothalamictemperature.
Variable
' Internal
body
AIR
40
AIR TEMP. 25.4 •
TEMP. 8.7 "C
AIR
34.5 'C
_^_R^CTAL^
-
RECTAL
38
TEMP.
36
skin"*"'"'"'""
34
•0
SKIN
32
30
28
26
•
-
,HEAT LOSS
X
oXo
3.5
x-HEAT
3.0
LOSS
*-HEAT LOSS
•-HEAT PROD.
•-HEAT PROD.
.
•'HEAT PROD.
x2.5
O2.0
.
^ 1.5
*
<
o
ic 1.0
RESPIR.Jj^T LOSS
•-RESPIR. HEAT LOSS
—RESPIR. HEAT LOSS
.^-PERSPIR. HEAT LOSS
— PERSPIR. HEAT LOSS
I
2
HOURS
3
5
0
12
HOURS
3
4
5
0
2
3-
HOURS
Rg- 2—Heat production, total heat loss, evaporative heat loss from mouth and from skin surfaces
averaged over half hour periods, and rectal and average skin temperature for a resting dog in a
cold, neutral, and warm environment. (Hammel, H. t., C. H. Wyndham and J. D. Hardy, 1958.)
They have been particularly concerned with quantitative measurements of the
thermoregulatory responses which the animal makes to various internal and ex
ternal thermal stresses. This information is essential for the rational manage
ment of economically productive animals.
For other less practical reasons, we, too, have investigated by indirect and
direct calorimetry the thermoregulatory responses of the resting dog to a range
of environmental temperatures. Those responses of the dog which could be mea
sured calorimetrically were plotted as a function of the time exposed to a hot,
a neutral, and a cold environment (Figure 2).
In Figure 3, the steady state responses are plotted as a-fur^6n~~Df^xrcrnal
environmental temperature. Here again these data are pertaining to theright side'
of the model in Figure 1. They may provide some informatioirabout the con
trolled system; about how its transfer functions relate the manipulated variables
and the disturbing signals to the regulated variable, that is to say, the thermo
regulatory responses and the thermal stresses to the internal body temperature.
40
RECTAL
38
'C 56
54
32
r
<5
30
x-HEAT
70
LOSS
* Ao
60
en
<mv. 50
8
8f»
8 *«
o
'4
d< 40
o
^
30
»-HEAT PROD.
20
K
1.2
V.
^^
<
o
ft
o
32
A
RESPIR. HEAT
LOSS
— *
28
.0
o
S"//-
>
24
r"N
ia"
°®
20
\
16
CONDUCTANCE
X
:a
\
12
.2 oo
6
COOLING
CONSTANT
rf>.« •
10
15
2 0
25
3 0
35
CALORIMETER TEMP. °C
Fig. 3—Equilibrium values of heat loss, heat production, evaporative heat loss from mouth, rectal,
and average skin temperatures, tissue conductance, and cooling coefficient of fur and overlying
air over a range of environmental temperatures from 10° to 35° C. (Hammel, H. T., C. H. Wyndham and J. D. Hardy, 1958.)
At the very least, some ofthe thermoregulatory responses have been identified
and quantified by these studies. Shivering is identified as oxygen consumption
and equated to heat production. Panting is identified as respiratory evaporative
heat loss. Variable peripheral blood flow is identified by its eflfect upon the ther
mal conductance of the tissue between the core and the skin surface. Heat trans
fer from the external body surface is identified as the difference between total
heat loss and respiratory evaporative heat loss. The rate of change of heat content may be identified as the difference between heat loss and heat production.
Air temperature is identified as one of the external thermal stresses and rectal
temperature is indicated as one ofthe internal body temperatures and therefore
a possible candidate for the regulated variable. The graphs provide some of the
transfer functions that are involved within the controlled system.
This aspect of temperature regulation still demands further exploration to
derive more sophisticated and detailed data and relationships describing (1) heat
content of components of the body, (2) heat transfer characteristics from one
component to another, (3) heat transfer characteristics from each part of the
body surface to the environment, as well as (4) parameters of the physical en
vironment.
I would like now to turn our attention to the other halfof the model de
picted in Figure 1, the controlling system.
Methods
Slowly we realized that little can be learned about the controlling system
from the type of steady state data represented in Figure 3. Although we intui
tively presumed that temperature sensing elements in the skin surface and in the
core ofthe body somehow generate the regulatory responses, we could learn
nothing about this process from the skin and rectal temperatures and the cor
related steady state responses as we had plotted them. We realized that we had
to use some technique for dissociating the several sensory inputs to the control
ling system from one another so that one sensory input at a time could be varied
and related to the regulatory responses while all other inputs remained unchanged.
Benzinger and his co-workers discovered one technique for achieving apar
tial dissociation of body temperatures in his extensive investigations on man
(Benzinger, Kitzinger, and Pratt, 1963). By taking advantage ofthe natural or
induced variations in three inter-related parameters, the evaporative heat loss by
sweating, the tympanic membrane temperature, and the average skin tempera
ture, and then extracting from these data aset of data (sweat rate vs. tympanic
temperature) for oneskin temperature and another set ofdata for another skin
temperature, and so forth, over a range ofskin temperatures, they were able to
plot the sweating rate as afunction of tympanic membrane temperature for each
of several average skin temperatures even though the environmental temperatvire
was in steady state. Unfortunately, for their results on sweating, they had to in-
duce variation in internal temperature by introducing exercise as a disturbing
signal.
To relate the shivering response to tympanic membrane temperature, Ben-
zinger et al, immersed their subjects in stirred water, achieving thereby a con
stant skin-temperature, and again took advantage of natural and induced varia
tions in the tympanic temperature.
Stolwijk and Hardy (1966) have used another dissociating technique. They
monitor the metabolic rate and sweating rate continuously both before and after
a step transition from one environmental temperature to another. Thus, in these
investigations on man, it has been possible to dissociate skin temperature from
internal temperature but, so far, it has not been possible in man to dissociate
the effects of one internal sensory site from another. It would be especially de
sirable to dissociate the hypothalamic temperature from the other internal tem
peratures where sensory inputs are almost certainly generated and fed into the
controlling system.
Concurrently, and over the past few years, we, too, have attempted to dis
sociate the sensory inputs by controllably manipulating the hypothalamic tem
perature by artificial means while recording the thermoregulatory responses in
the conscious and wakeful dog. Figure 4 depicts the experimental system into
FRAME
V4
V9
CAS04!=j H0SO4
OUTDOOR
CALIB.
GAS
PUMP =
SILICA GEL
H.SO
COND.
DC
AMPLIFIER
BUCK
CASO4
RECTAL T
POT
02
HYPO
AVE. SKIN
[ANALYZE R
T
AIR
T
EAR
T
TRUNK
TRUNK
0,
ANALYZER
EXHAUST
AIR
T
WET GAS METER
RECORDER
Fig. 4—System for recording oxygen consumption, evaporative heat loss from mouth, rectal, skin
and hypothalamic temperatures, air temperature and temperature of stimulating water In head
circular. (Hammel, H. T., D. C. Jackson, J. A. J. Stolwijkand J. D. Hardy, 1963.)
Thermocouple
,
^0 a From Constant Temperature Both
To
Vacuum
Thin wall stainless
Thermode
55 mm x 20 gauge
Thin wall stainless
Catheter
Fig. 5—Details of thermode (or re-entrant tube) and circulator construction. Thecirculator is shown
In place for thermal stimulation of the hypothalamus. Only the lower acrylic plate or bottom of
the circulator is left permanently attached to thethermodes and guides byepoxy resin. (Hammel,
H.T., D. C. Jackson, J. A.J. Stolwljk and J, D. Hardy, 1963.)
which the dog is introduced for monitoring its regulatory responses. Long and
arduous training is required to train the dog to accept the experimental proce
dure without discernible emotional reaction even though there is never any pain
or even discomfort associated with the experience. Up to two years prior to ex
perimentation, the dogs were implanted with seven thermodes surrounding the
pre-optic area and anterior hypothalamus as depicted in Figure 5. About 150
ml/min of water at any selected temperature perfuse the thermodes during an
experimental period. The small plastic circulator with its thin wall stainless cath
eters (0.015 "ID and 0.019" OD) projecting into the thermodes is shown
mounted above the thermodes and on the head. In addition, a re-entrant tube
passes through the brain to the anterior commissure and 1 mm from the mid-
line. Athermocouple is introduced into this tube to monitor the temperature of
the hypothalamus. A cross section through the brain stem at a level of 25.5 mm
anterior to the stereotaxic ear bars. Figure 6, clearly shows how the pre-optic
tissue between the optic chiasm and anterior commissure is straddled and sur
rounded by thermodes. Although the damage to the hypothalamic tissue is min
imal and negligible after a recovery period of one week, there is an obvious
Fig. 6—Section through the brain of dog at 25.5 mm anterior to stereotaxic ear bars showing AC,
anterior commissure; OC, optic chiasm; T25, thermode at 25.0 mm; T22, thermode at 22.0 mm;
and R, re-entrant tube at 26.5 mm, left 1.0 mm from midline.
limitation to this method of thermal stimulation of the hypothalamic tissue.
Heat is transferred to (and from) the tissue from (and to) the thermodes by con
duction through the tissue and by the blood perfusing this highly vascular area.
There are, therefore, unavoidable temperature gradients throughout the tissue
whose response is investigated. Furthermore, since we are unable to control the
position of the thermodes with respect to the major vessels subserving the preoptic region, we cannot predict the pattern of the isotherms through the tissue,
nor can we say how the one temperature measured in the re-entrant tube relates
to the average pre-optic temperature. I have dwelt briefly on this technical limi
tation, since I shall have to refer to it again in order to explain some bizarre re-
11
sponses observed when the pre-optic temperature is greatly displaced (as much
as 5° C) from the internal blood temperature.
The Regulated Temperature
In an inconspicuous way, I have introduced one major assumption which
requires further comment. I have inferred that the temperature of the hypothaia
mus is the regulated internal body temperature. Actually, I can only expect to
convince you by direct evidence that it is an important temperature input to
the controlling system. That the hypothalamic temperature is the regulated tem
perature of the body can only be inferred from its relationships to other body
temperatures and the thermoregulatory responses.
Any recent review oftemperature regulation (Hardy, 1961; von Euler, 1961)
is replete with evidence that sizeable displacements of the hypothalamic tempera
ture (3° to 5® C) will activate panting or shivering in many species (unanesthe-
tized); and ifthe displacement is great enough (5° to 20° C), Andersson's group
have shown that neuro-humoral responses are activated with obvious thermal
consequences (Andersson, Gale, and Hokfelt, 1964). In a way. Figure 7 may
serve to summarize the evidence that the rostral hypothaiamus is sensitive and
responsive to small displacements ofits own temperature. The widely swinging
•c
«r-
Botk Braltt
RlCtlin
4m(ti.cyclM
ISinlfucycUt
9m«i.eyclM
ESSnSTcycBs
lOmln. cyeitt
l9niIn.circlM
cyeitt
clr€.itop.
Fig. 7—Cyclic heating and cooling ofhypothaiamus with water perfusing thermodes alternately at
41.0° C and 35.0° C, Ambient temperature was 25' C. The 30 minute cycle equals 15 minutes
heating and 15 minutes cooling. (Hammel, H. T., S. B. Stromme, and R. W. Cornew, 1963.)
12
temperature (from 36.0° to 40.5° C) in this record is that ofa thermocouple in
the middle of the anterior hypothalamus while the hypothalamus is alternately
heated and cooled at several frequencies. The low-amplitude traces in the middle
of the widely swinging trace are those of the cerebellar ("back brain" in Figure
7) and rectal temperatures swinging in consonance with the hypothalamic tem
perature. Above these temperature traces is the trace of the ear pinna tempera
ture of the dog in an ambient temperature of 25° C. When the hypothalamus
was alternately cooled and heated, the blood vessels of the ear pinna were al
ternately constricted and dilated, resulting in an alternately falling and rising
ear pinna temperature. Likewise, the rectal and cerebellar temperatures were al
ternately rising and falling at the same rate and by about the same magnitude
while the hypothalamic temperature was displaced by equal amounts below or
above its unperturbed level.
We conclude from this that the anterior hypothalamus is responsive to both
heating and cooling and with equal sensitivity to moderate heating and cooling.
We may also note that the response time of the controlling system is less than
one minute, i.e., within a minute after the hypothalamic temperature decreased,
the ear pinna temperature begins to decrease and begins to increase after the
hypothalamic temperature is increased. This example illustrates that the control
ling system, as well as the system which is controlled, namely the vasomotor re
sponse, possess a rapid response time. The example also illustrates that the rate
of change of internal temperature following a step change in hypothalamic tem
perature is greatest initially and gradually diminishes 10 to 15 minutes after the
hypothalamic temperature change.
If we are convinced that induced deviations of the hypothalamic temperature
by a few degrees above and below some normal value do activate appropriate
thermoregulatory responses, we should then wonder how much the hypothalamic
temperature departs from the norm during a natural thermal stress when these
same regulatory responses are activated. In a neutral environment, the hypo
thalamic temperature is variable in a resting dog; in an average dog, it may
range from 38.1° to 38.6° C during a period of a few hours. It is also variable
in a hot environment and in a cold environment and in a way which is not dis
tinguishable from the variability in the neutral environment. My colleague, Dr.
B. Hellstrom, has found that the range of the hypothalamic temperature, the
mean hypothalamic temperature and its standard deviation are not different in
the same dog when in a hot, a neutral, or a cold environment, even though pant
ing is vigorous in the hot environment (40° C) or shivering is vigorous in the
cold environment (10° C) and there is neither panting nor shivering in the neu
tral environment (25° C). Of course, the hypothalamic temperature can be forced
to deviate from its normal range by a stress so severe as to produce a near maxi
mum response.
To summarize our findings so far, we believe we have established the pre-
optic region rostral to the hypothalamus to be a temperature-sensitive part of
.13
the brain stem responding to moderate heating by activating heat dissipation
and to moderate cooling by activating both heat conservation and increased heat
production. Moderate heating and cooling may mean only 1 or 2 or 3° C above
or below 38.5° C, yet we never saw deviations in the hypothalamic temperature
of this magnitude when the resting animal was exposed to a range ofexternal
thermal stresses. In fact, no differences in hypothalamic temperature were pro
duced by the environmental temperature.
Properties of the Controlling System
We now must recognize two essential features of any regulated system;
these are depicted in Figure 1. First, the output of the controlled system must
somehow be fed back as a signal to the controlling system. In temperature reg
ulation, the consequence of the thermoregulatory responses and the thermal
stresses acting upon the controlled system yield, after appropriate transforma
tions, an output which is the internal body temperature. The output, or some
part of it, the hypothalamic temperature, is fed back to the controlling system.
Second, within the controlling system there is always some provision for gener
ating a reference signal with which the feedback signal is compared and from
which an activating signal is derived. The final consequence of the activating sig
nal is such as to reduce the difference between the feedback and reference signals.
Recognizing that the hypothalamic temperature has no significance by itself
and meaning is given to it only when there exists some mechanism for compar
ing it with a reference signal and from which a difference or an error signal can
be generated which can activate a corrective response, we are now led to wonder
about the nature of the reference signal. Is the reference signal, often called the
set temperature, generated within the hypothalamus (or its associated preoptic
area) so that this area by itselfcould regulate body temperature without addi
tional inputs.^ Is the reference signal or set temperature invariant and uneffected
by other known inputs to the hypothalamus or is it modified, modulated, ad
justed by some or all of the inputs to the hypothalamus.' What are the relation
ships between the activating signal and the thermoregulatory responses; are they
linear.' Are the relationships modified by other inputs to the hypothalamus?
Of course, these questions would be answered if we knew how the neurons
in the hypothalamus and the preoptic area are interconnected and function. How
ever, we do not know the neuron circuitry for the controlling system for tem
perature regulation; therefore, we are compelled to seek experimental evidence
relevant to the questions and then perhaps make some reasonable guesses about
the neuron circuitry.
We have undertaken to explore these questions in a systematic way in the
dog and other experimental animals. We are using two approaches which yield
complementary information. Our first approach is illustrated in Figure 8. For
this record, the dog was awake and resting in a neutral environment and was
Ecr Ptnno
2 40
Roctum
1
No circ.
Fig. 8—Heat production, respiratory heat loss, and skin and rectal temperatures in response to
heating and cooling hypothalamus. Airtemperature = 23.0 ± 0.5° C from 0 to 273 min. Numbers
over level segments of hypothalamic temperature indicate temperature of water circulating In
thermodes. (Hammel, H. T., S. B. Stromme and R. W. Cornew, 1963.)
15
isolated in an environmental chamber. Its hypothalamic temperature was ma
nipulated to several levels above and below normal for approximately 10 minutes
each by perfusing the thermodes with water at the appropriate temperature
(noted in Figure 8 by the number above the level segment ofthe record ofhypo
thalamic temperature). The oxygen consumption and the evaporative water loss
of the dog were recorded as heat production and respiratory heat loss in Kcal
kg'^ hf\ Ear pinna, neck skin and rectal temperatures were also recorded. Since
we believe that internal body temperatures (other than hypothalamic) are also
sensed and provide input signals to the controlling system, each displacement
of hypothalamic temperature from normal was initiated only after the rectal tem
perature returned to 38.1° C. For the same reason, each period of hypothalamic
temperature displacement was kept short; any strongly activated regulatory re
sponse would markedly alter the internal body temperature when applied for
even 10 minutes and this altered internal body temperature would, in turn, effect
the activating signal.
The results shown in Figure 8 are utilized in Figure 9 where, for example,
in the neutral (23° C) environment the heat production for a 3-minute interval
(including the peak response) was plotted as a function of hypothalamic tem
perature. From this curve, two important characteristics ofthe regulating system
may be determined for the metabolic response: (1) The threshold hypothalamic
temperature above which only basal or resting metabolism was obtained and
34
36
38
40
42
HYPO. TEMP.
Fig. 9—Heat production as a function of hypothalamic temperatures for a quiet, resting, wakeful
dog at three air temperatures, 13.5° C, 23° C, and 33.5° C. (Hammel, H. T., S. B. Stromme and
R. W. Cornew, 1963.)
16
below which the metabolism of shivering was added and, (2) the shape of the
shivering response curve for hypothalamic temperatures below threshold. For
this wakeful, resting dog in a neutral environment, the threshold temperature
was approximately 37° C or about 1.5® C below the unperturbed hypothalamic
temperature. The shape of the response curve below threshold may be roughly
approximated by a straight line with a slope of about 2 Kcal kg"^ hr'^ C'\
Repeating the same procedure that was used for Figure 8 and in the same
wakeful, resting dog but in a cool environment (13° C), another set of results
were obtained which arc shown in Figure 9. In the cool environment, the thresh
old hypothalamic temperature for shivering has now increased to about 39° C.
The normal range of hypothalamic temperatures is below the threshold level so
that the dog shivers a little even without any artificial displacement of tempera
ture below normal. The shape of the response curve was roughly the same as
was found in the neutral environment and the slope was about 2 Kcal kg"^ hr*^
c-\
The threshold for panting was approximately 39° C in the neutral environ
ment and about 41° C in the cold environment. The slope of the curve for the
panting response was estimated to be between l*ancl 5 Kcal kg"* hr"' C'\ My
colleague. Dr. Hellstrom, has thoroughly investigated the thresholds and slopes
of the response curves for both shivering and panting in hot, neutral, and cold
environments for dogs in the resting and wakeful state. Although his results
(still unpublished) are much more extensive, they are similar in magnitude and
shape to those reported here. Even though he has many more data, the variabil
ity in the activated regulatory response for a given hypothalamic temperature is
no less than is evident in Figure 9, suggesting that we were either unable to
control all other inputs to the controlling system or that the controlling system
is inherently noisy.
Our second experimental approach for obtaining the characteristics of the
controlling system is illustrated in Figure 10. In these experiments, the normal
combination of hypothalamic, internal body, and skin temperatures is dissociated
by prolonged displacement of the hypothalamic temperature. This leads to hypo
thermia or hyperthermia in the other internal body temperatures, the magnitude
depending upon the environmental temperature.
In the experiment of Figure 10, the hypothalamus was heated, during which
time the dog was in a neutral and then in a cold environment. The heat content
of the body was markedly reduced as indicated by a decrease in rectal tempera
ture, a decrease greater than 1° C. Upon release of the thermal clamp on the
hypothalamus, the hypothalamic temperature fell approximately 2° C to assume
a value slightly above the hypothermic rectal temperature. The shivering present
just prior to clamp release (and which could only be attributed to some nonhypothalamic, internal body temperatures providing inputs which react in some
way with the hypothalamus) was followed by a large increase in metabolism
(nearly four times the resting level) just after the clamp release. Assuming that
17
.IsJtrurW
Ts(oar)
RESR/MtN.
^
Lxgctol
CoL/kg/hr
SO
^/VaTVvA
Metobolism
Minutes «
Fig. 10—Body temperatures and heat production of resting, fasting dog exposed to neutral and
cold environments during monipulatlon of its hypothalamic temperature. (Hammel, H. T., D. C.
Jackson, J. A. J. Stoiwijk, J. D. Hardy, and S. B. Stromme, 1963.)
no changes in the internal body temperature or skin temperatures occurred dur
ing the brief interval that the metabolism and hypothalamic temperature were
changing from one level to the other and assuming that no bther inputs to the
hypothalamus changed so that the threshold for shivering was the same before
and after clamp release, we can then calculate from the shivering metabolism and
its associated hypothalamic temperature both before and after clamp release what
the slope of a linear response curve would be. In this test, the slope was found
to be 2.0 Kcal kg ' hr'^ C"*. After the slope is known, the threshold during the
interval when we are assuming that it is not changing may also be calculated.
It was 39.6® C in this test.
I have described two procedures whereby we can dissociate the hypothalamic
temperature from the other thermal inputs to the controlling system in experi
mental animals and from these results determine at least roughly the threshold
values and the shapes of the response curves for any thermoregulatory response
under any set of experimental conditions. Eventually, we expect to obtain this
information for suitable experimental animals under the following conditions:
at rest and in exercise; awake and asleep; exposed to hot, neutral, and cold en
vironments; in the normal and the fevered state; during hibernation and out;
during emotional stress; during pharmacological stress; cold and heat acclimated,
etc.
18
Adjustment of the Set Point
I would next like to anticipate some of the results to be obtained from the
systematic investigation of the multifarious aspects of body temperature regula
tion I have just outlined. To do so, I will describe briefly some preliminary ex
periments which served to give direction to our experimental program.
The effects of sleep upon body temperature regulation may be anticipated
by the results shown in Figure 11. The hypothalamic temperature of a rhesus
monkey restrained in a primate chair was recorded during a 24-hour period of
exposure to a hot (35° C), a neutral (30° C), and a cool (20° C) environment.
At 1800 each day, the lights in the environmental chamber were turned off and
at 0900 the next day they were on again, although daylight entered through the
chamber window after 0600.
At each environmental temperature, it appears as if the thresholds for thermoregulatory responses were reduced byoneor more degrees C during the dark period
when, presumably, the monkey was less active and asleep. At first, we mightsup
pose that the fall in brain temperature at the onset of sleep in the neutral and
cool environments could be due to a reduced level of heat production combined
with a decrease in the slope of the thermoregulatory response curves. However,
if the slope of the response curves had diminished at the onset of sleep, then
in the hot environment the brain temperature would have increased to a higher
isSs
M^o—i^oo—nbe—rySs—siss—nbs—i^ss——^59—is^s
Sioo
5foo
0500
"OMO
oioo
0700
TIME (EOT)
0S06"
1000
i
1100
Fig. 11—Hypofhalamic temperatures of a rhesus monkey restrained in a primate chair in hot (35°
C), neutral (30° C), and cold (20° C) environments (50 percent relative humidity) for 24 hour pe
riods with normal day-night lighting. (Hommel, H. T., D. C. Jackson, J. A. J. Stolwijk, J. D. Hardy
and S. B. Stromme, 1963.)
19
level during sleep in order to activate an increased heat loss to balance the heat
production. In fact, the brain temperature decreased even in the hot environ
ment. Figure 12 may be interpreted in the same way.
For this record, the ear pinna and hypothalamic temperatures were related
to the eyelids in a rhesus monkey in a cool environment for 7 hours. On every
occasion when the monkey closed its eyes, its ear pinna temperature increased
even though its hypothalamic temperature was decreasing at the same time. Con
versely, when the animal opened its eyes, its ear pinna temperature fell while
its hypothalamic temperature was increasing. Increase or decrease in activity,
when the eyes were open, did not affect hypothalamic or ear pinna temperatures.
These results suggest that the thresholds for thermoregulatory responses are de
creased at the onset ofsleep and increased upon awakening. They do not sug
gest that the slopes of the response curves are effected by sleep.
Even more convincing are observations of the increase in sweat rate in man
at the onset ofsleep obtained by Kosuge in 1936 and discussed by Kuno (1956).
In one series of experiments at 19° to 20° C in the winter, subjects did not
sweat on the chest before or during sleep. At 29° C, a subject did not sweat be
fore sleeping but within 30 minutes of the onset ofsleep the subject began to
sweat moderately on the chest and stopped as soon as he was awakened. In
Ear Pinna
rEyes open
Eyes closed
r Act ve
Inactive
3
4
HOURS
Fig. 12—Hypothalamic and ear pinna temperatures of a rhesus monkey in a primate chair in cool
environment {22»-24° C}. Open or closed eyes and activity are noted. (Hammel, H. T., D. C. Jack
son, J. A. J. Stolwijk, J. D. Hardyand S. B. Stromme, 1963.)
20
other experiments at 29® to 32° C in the summer, the waking subjects sweated
more or less (77 to 140 gm per hour). Sweating increased at once, after they
fell asleep, and varied from 95 to 180 gm per hour. It subsided immediately
when the subject was awakened. Kuno assumed that the increase in sweating
was due to an increase in excitability of the thermal sweat center resulting from
a reduced tonic inhibitory action from an inhibitory center in the cerebral cortex.
He supposed that the increased excitability of the sweat center was due to an
increased sensitivity; i.e., a greater response was obtained from the same stimu
lus. I would reinterpret these observations by suggesting that the sensitivity
(slope of the response curve) does not increase at the onset of sleep, but that
the set temperature (threshold) decreases at the onset of sleep and increases
again upon awakening.
In a preliminary investigation of the regulation of body temperature during
exercise in the dog, we concluded that the set temperature was decreased at the
onset of exercise, other conditions remaining the same. This conclusion is con
trary to the interpretation which Nielsen placed upon his data in man (Nielsen,
1938), from which he decided that the set temperature increased during exercise.
He observed that the rectal temperature always increased to the same level for a
given work load regardless of the environmental temperature. Although Niel
sen's observations have been confirmed many times, and his interpretation has
been repeatedly subscribed to by subsequent investigators, it seems unreasonable
to us to suppose that the set temperature should increase. An elevated set tem
perature would deprive the regulator of a fraction of its activating signal and
would require the regulated temperature to climb to still higher levels in order
to generate a signal which could activate the required and greatly elevated heat
loss during exercise. At least, the set temperature should not increase and at
best it should decrease at the onset of exercise, thereby generating an immediate
signal to activate increased heat loss in anticipation of the need to dissipate the
increased heat production. In Figure 13 is a summary of the evaporative heat
loss by panting in three exercising dogs (running at 4 mph) plotted as a func
tion of the hypothalamic temperature. The evaporative heat loss for the same
dogs while resting just prior to running is also shown. In order to obtain a
range of hypothalamic temperatures, dogs were run in hot, neutral, and cold
environments. The slope of the response curve during exercise was about 6 Kcal
kg'^ hr"^ C'\ However, this slope is probably greater than would be obtained
if the environmental temperature had been the same throughout. For the higher
hypothalamic temperatures in Figure 13, the air temperature was 30® C so that
the rectal temperature was a few tenths of a degree higher than was found in
the neutral environment and the skin temperature was 1° to 2° C higher. There
fore the evaporative heat loss for the highest hypothalamic temperatures was
higher than it would have been in a neutral environment. For the same reasons,
the evaporative heat loss for the lowest hypothalamic temperatures (obtained at
an air temperature of 15° C) was lower than it would have been in a neutral en-
21
10
9
EVAR 8
HEAT^
LOSS
KC^
6
KG 5
DOG
B
C
D
Exercise
•
a
.
Rest
o
A
•
I
i__|
oa6a 8!
OB
o
r
38.0
§"0- A-°—Oj A
J
1
.5
1
X
1
1
I
39.0
I
I
.5
HYPO. TEMP. (®C)
Fig. 13—Evaporative heat loss as a function of hypothalomic temperature at rest and In exercise.
(Jackson, D. C. and H. T. Hammel, 1963.)
vironment. Although we cannot ascertain precisely what the slope would be
during exercise in a neutral environment, we may readily see that the response
curve would extrapolate to a threshold level of panting at a hypothalamic tem
perature below 38° C, possibly 37° C or even lower. Since the threshold hypo
thalamic temperature for panting in the resting dog in a neutral environment
is about 39° C, the set temperature while running at 4 mph was approximately
2° C below the resting set temperature. It is certain that the threshold tempera
ture for panting did not increase during exercise even though the hypothalamic
temperature did increase.
The set temperature appears to shift downward immediately with the onset
of exercise and to shift upward again when exercise stops. This deduction may
be made from Figure 14. The measured evaporative heatloss bypanting increased
immediately (limited only by the response time of the measuring system) with
22
Kcal/kg/hr
Evap. Heat
®C
Loss
40.0
10
8
6
39.0
4
hypo.
2
0
38.0
0
15
30
45
60
M inutes
Fig. 14—Continuous evaporative heat loss recording on Dog D at 30° C. Exercise at 4 miles per
hour is indicated by dotted line. Note sudden changes in evaporative heat loss associated with
the beginning and end of the exercise period. (D. C. Jackson, 1963.)
the onset of exercise and before any significant change in hypothalamic tempera
ture occurred. The panting, of course, continues to increase as the hypothalamic
gradually increases. When exercise stopped, there was an immediate decrease in
evaporative heat loss, although it did not decrease to its normal resting level in
a 30° C environment since the hypothalamic was still elevated well above its
normal range. Only gradually did panting diminish as did the hypothalamic
temperature.
In 1936, Iwatake made parallel observations in man which are discussed by
Kuno (1956). Sweating increased greatly above a low pre-running sweat rate and
within minutes of the onset of running, before any significant increase in rectal
temperature was observed. The rate of increase and the profuseness of the sweat
ing was dependent upon the speed of running. At the end of running, there
was an immediate decrease in sweat rate with no change in rectal temperature
followed by a gradual decline in sweat rate. Again, a decrease in the set tem
perature at the onset of exercise and an increase in the set temperature at the
end of exercise seems to describe the changes in the regulatory system control
ling sweating in exercising man.
Next, I would like to consider the cortical input into the system which regu
lates body temperature. We all arefamiliar with our own ability to suppress volun
tarily the shivering response at least for a short time even in a very cold en
vironment. We also can simulate shivering in a neutral or hot environment.
Presumably, the dog can do the same thing, as indicated in Figure 15. Heat pro
duced by shivering in a cold environment was interrupted for about 10 minutes
in these three examples, so that the metabolic rate dropped to the resting level
(broken line). Because of the reduced heat production, the internal body tem
perature fell 0.1° to 0.2° C. We could not argue the converse; the heat produc
tion did not drop to the resting level because the internal body temperature
23
RECTAL
.6r
CAL.TEMR ll.3*C
HEAT PROD.
ALCOHOL LAMP
K
z
?
-4
3
1
j
.2
<
DOG B
o
MAY 2,56
10
15
20
MINUTES
25
°4:5
30
HOURS
CAL.TEMR 8.6*0
RECTAL
.^,37.8p ,A , RECTAL
5.0
CAL.TEMR 16.9 *0
37.6 -
HEAT PROD.
.HEAT PROD.
K
X
^5 2
DOG A
NOV. 15, 55
2.5
HOURS
DOG C
JUNE 26.56
3.0
2.5
HOURS
3.0
Fig. 15—Recordsof oxygen consumption and rectal temperature in three dogs exposed to cold
environment. The records were selected to cover the brief period when the shivering metabolism
was voluntarily suppressed so that heat production dropped to the resting level indicated by the
broken line. A record of firing of one alcohol lamp and a recordof two burning alcohol lamps with
a brief switch to outdoor air indicate the response time of the Oj meter.
overshot the threshold temperature and turned off the shivering. Shivering did
not start again because the internal temperature dropped well below the thresh
old. Part of the overshoot in heat production in Dog A may have been due to
the marked drop in rectal temperature during the period of suppressed shivering.
In a similar way we may presume that the dog can voluntarily interrupt panting
for short periods in a hot environment.
Of course, we cannot say how the cortical input enters the controlling sys
tem, but it acts as if the threshold is lowered to suppress shivering voluntarily
in a cold environment or is raised to suppress panting in a hot environment.
Similarly, shivering may be voluntarily activated in a hot environment by raising
the threshold to a level above the actual hypothalamic temperature. Whether the
threshold for panting (or sweating in man) is also raised by the same amount
by the cortical input is not known. If it were to increase, then panting (or sweat-
24
ing) should be suppressed while shivering is activated. Alternatively, the cortical
input may enter the final pathway for the activated regulatory response after the
activating signal has been generated in the hypothalamus. Then only a single
response would be affected voluntarily.
If I were a cautious experimentalist, I would wisely terminate this discussion
of temperature regulation with a summary of our experimental evidence. The
evidence suggests to us: (A) The pre-optic region associated with the hypothala
mus is responsive to its own temperature and when its temperature is displaced,
thermoregulatory responses are activated. (B) The hypothalamic temperature by
itself cannot aaivate a regulatory response. An activating signal is generated only
when the hypothalamic temperature is compared with a reference signal (which
may be identified as the threshold temperature for a given regulatory response).
(C) The magnitude of the regulatory response is a function of the activating
signal; the latter is taken to be the difference between the hypothalamic tem
perature and the reference (or threshold) temperature and the function are de
fined by the shape of the response curve. (D) The response curve may be linear
with a slope which is characteristic for the response. (E) The sign of the activat
ing signal determines which response is activated. The appropriate responses
are those which tend to reduce the magnitude of the activating signal. (F) En
vironmental thermal stress, internal heat stress, and other known inputs to the
controlling system act as if to adjust the reference input signal. A cold stress
acts to raise the thresholds for all regulatory responses, a heat stress acts to lower
the thresholds; exercise acts to lower the thresholds; so also does sleep lower
the thresholds, etc.
I should stop at this point. However, I wish to continue in order to (1)
formalize these summary statements, and (2) suggest a model of interconnected
neurons which could conceivably perform the function of the controlling system
for temperature regulation.
The Law of the Controlling System
An important step in elucidating the nature of a controlling system is to
obtain the so-called "law of the controlling system," that is, the relationship
between inputs and outputs. All of the above summary statements arc described
by an equation of the form
R ~ Ro —
(Thypo ~ Tget ) >R ~
O
R
where R - Rq is the thermoregulatory response (metabolism, vasomotor aaivity,
sweating, panting, behavior, and so forth);
is the basal level when T^ypo =
Tset ; Thypo, the actual hypothalamic temperature, is the feedback signal; T^et ,
R
R
the functional set point (or threshold) temperature for response R, is the refer
ence input signal; (T^ypo - Tget ) is the actuating signal; and Or is the proporR
25
tionality constant for the response (R - Ro).
This equation states that a given thermoregulatory response is proportional
to an actuating signal which is the difference between the actual hypothalamic
temperature and some threshold temperature for that response. The equation
formalizes summary statements (a) through (e). Summary statement (f) may be
formalized by stating that the set (or threshold) temperature for response R is a
fiinaion of all the inputs into the controlling system. Perhaps the simplest func
tion would be a sum of constant and variable terms, each term representing the
effect of each of the inputs. Thus,
Tset = an intrinsic (and constant) hypothalamic set point
R
±
+
-
a constantoffset term (+ for panting and - for shivering)
f (skincold receptors)
g (skin warm receptors)
h (non-hypothalamic core receptors)
a sleepterm (possibly graded)
+ a fever term
- an exercise term (possibly graded)
±
a cortical term
The second or "offset" term was introduced to account for the fact that the
threshold or set point temperature for panting is higher than that for shivering.
The first two terms, the "intrinsic" and the "offset" terms, are presumably gener
ated within the pre-optic area so the hypothalamus could function as an auton
omous controlling system requiring no additional inputs and generating a regu
latory response simply by the deviations of the hypothalamic temperature from
the fixed set temperature.
To characterize the other functional terms in any quantitative way is im
possible at present. It may be instructive, however, to utilize what is known
about thermal receptors within the body in order to discern how these terms
may vary. We may reasonably assume that the functional term for the skin cold
receptors is proportional to the firing rate of the cold receptors. Limiting our
attention to the initial and near linear portion of the steady state discharge rate
of the "cold" receptors as obtained by Hensel et ai, in the cat tongue (1951),
cat furred surface (I960), and human skin (I960), we may say that the firing rate
is proportional to (40 - Tgkin). i-e*, as the skin temperature decreases below 40°
C the discharge rate increases about linearly down to about 20° C. If the skin
temperature is changing in time, then of course there is also a phasic discharge
rate. We may guess, therefore, that
f (skin cold receptors) =
(40 - Tg) -f ^ Tg.
In a similar way we may guess that
g (skin warm receptors) = y(Ts - 37) -t- 7 Tg
26
where
y and y are unknown proportionality constants. Although we have
ample evidence to presume that there are thermal receptors within the core of
the body in addition to those within the hypothalamus, these have not been
investigated in a way to relate discharge rate to temperature; therefore we shall
guess only that
h (non-hypothalamic core receptors) = 5 Ts .b.
where Ti
is the internal body temperature.
Ignoring the other inputs, such as exercise, sleep, etc. and considering only
the interactions due to external theripal stress as this stress affects skin and in
ternal body temperatures, we can combine the.above terms into the controlling
equation and obtain,
R - R„ = ttR Thypo + ATr - (40 - T^) - ^8 T,.
+ Y (Ts - 37) + 7 Tg + 5 Tj b.
Ignoring the possibility that our several simplifying assumptions may need cor
rection, and ignoring the possibility that a measurement of the average skin tem
perature is not properly weighted for its effect upon the controlling system, the
above equation is still of little value in obtaining a quantitative estimate of the
thermoregulatory response since there are at least five unknown constants. The
equations indicate what we already know which is that, at least, the hypothalamic temperature, internal body temperature, skin temperature, and time rate of
change of skin temperature are involved in generating the response. At the
present time we are satisfied to express the effect of environmental temperature
as a measurable shift in the functional set temperature.*
♦
- Footnote added when editing the lecture
In considering the simplest function to be used for expressing the functional set temperature,
we chose to use a sumof terms. Equally simple would have been a product of terms, but thiscan
be easily ruled out. If any one of the terms is zero, then there would be no reference input signal
and, consequently, no regulation. The experimental evidence is strongly against this possibility.
We may wish to guess as suggested by Hardy and Stolwijk that the regulatory response is
also derived from a product of terms, from the hypothalamus, the skin, the internal body tempera
ture, etc. Thus, for example,
R - R„ = a„ (Thypo -T, ) (T, - T, )
O
R - R, > O.
O
This expression is also clearly forbidden. It states that if any of the terms including the hypothalamic term is zero or negative, there will be no regulatory response no matter what the magnitudes
of the other terms are. Gentle prolonged heating of the hypothalamus of a dog, which is shivering
in a cold environment, will suppress shivering initially; but as the internal body temperature falls,
shivering is restored to almost the same level, and the internal body temperature levels off to a
slightly lower level. Shivering is not indefinitely suppressed and profound hypothermia is not pro
duced by gentle warming of the hypothalamus. Similarly, gentle prolonged cooling of the hypo
thalamus of a dog which is panting in a hot environment docs not indefinitely suppress panting
and does not lead to fatal hyperthermia. The panting is initially suppressed but is restored within
a few minutes as the internal body temperature rises to a new and slightly higher level. The effect
is as if the increased internal body temperature dropped the set temperature for panting to below
the level to which the hypothalamus is being cooled so that panting is restored.
27
Ultimately, we would hope to be able to understand the relationships be
tween the inputs and outputs, the law of the controlling system, in terms of
the circuitry and activity of those neurons which are presumed to constitute the
controlling system. At present, any treatment of the controlling system in terms
of its neurons will be clearly recognized as guessing. Nevertheless, I shall at
tempt to make a guess (Hammel, 1965).
Neuron Model of Temperature Regulator
Our model will be based on the following assumptions:
1. There are neurons in the rostral hypothalamus having spontaneous firing
rates which are strongly temperature dependent, i.e., Qio >!> 1, over the range
of normal deep body temperatures. These are designated as hi-Qio primary sen
sory neurons.
2. Axons of these sensory neurons synapse with neurons within the hypo
thalamus which ultimately activate thermoregulatory responses. These latter are
designated as first stage or primary motor neurons.
3. The primary motor neurons may or may not have spontaneous firing
rates depending upon the choice of models to be preferred. Their firing rates
are assumed to have little or no temperature dependence except as influenced by
the sensory neurons.
4. Synaptic terminations on cell bodies of both primary sensory and primary
motor neurons may either facilitate or inhibit the neurons.
5. Although studies employing experimentally induced lesions and electrical
and thermal stimulation may indicate that primary sensory and motor neurons
are found in high concentrations in certain hypothalamic sites, these results can
not be interpreted to mean that neurons of a given type are located only in a
small, circumscribed region.
These are not unreasonable assumptions to make regarding neuronal activity
and are generally accepted as working assumptions on the limited evidence avail
able. An additional assumption will be made at this time; its justification, or
rather its desirability, will become apparent later.
6. We shall assume that another set of primary sensory neurons designated
as lo-Qio sensory neurons is located in the rostral hypothalamus in the same
region as the hi-Qio sensory neurons. The lo-Qm sensory neurons are assumed
to havespontaneous firing ratesoverthe rangeof deep body temperature. Further,
suppose that the cells are either not strongly temperature dependent, i.e.,
Qio —1, or they increase their firing rate with decreasing temperature, i.e.,
Qio <! <C 1- Like the hi-Q^o sensory neurons, the lo-Qio sensory neurons are
assumed to synapse with and facilitate or inhibit the action of the primary motor
neurons which activate regulatory responses.
In Figure 16, we are suggesting one way in which the hi-Qjo and lo-Qio'
sensory neurons are connected with three classes of primary motor neurons which
28
respectively activate panting, vasoconstriction, and shivering. In this figure we
have included two more assumptions, one essential and the other trivial. It is
essential to assume that the hi-Qio sensory neurons facilitate primary motor
neurons increasing heat loss, e.g., panting, and at the same time inhibit primary
motor neurons which lead to vasoconstriction and shivering. Conversely, the
lo-Qio sensory neurons must inhibit panting and at the same time facilitate
vasoconstriction and shivering. The other assumption is that the primary motor
neurons for panting and shivering in Figure 16 have no spontaneous firing rates.
To obtain different set-point temperatures for panting and shivering, we have
shown more inhibition than facilitation from the sensory neurons synapsing
with the neurons for panting and shivering. This provision generates the "offset"
term in our formal equation describing the functional set temperature. The same
condition could also be achieved by assuming the motor neurons to have a nega
tive bias so that more facilitation than inhibition would be required to activate
panting and shivering.
The activity curves of the sensory and motor neurons below the diagram
in Figure 16 indicate how the controlling system is presumed to function. First,
examine the set of activity curves for the neutral environment. For the tempera
ture at which the firing-rate curves of the hi-Qjo and lo-Qio sensory neurons
are equal, i.e., intersect, the facilitation and inhibition from these sensory neurons
upon the vasoconstriction (v.c.) motor neurons nullify each other so that the
v.c. motor neuron is active at its own spontaneous firing rate. The temperature
at which the firing-rate curves of the hi-Q,o and lo-Qio neurons intersect in a
neutral environment for a resting, wakeful, normal animal is designated to be
Th .** If the hypothalamic temperature drops below T^ in the neutral environ
ment, then the v.c. motor neuron is more facilitated than inhibited and vaso
constriction is increased. If the hypothalamic temperature rises above Th , then
n
vasoconstriction decreases. The temperature at which vasoconstriction becomes
zero is designated as the functional set-point temperature for vasoconstriction in
a neutral environment.
As shown in the activity curves for the neutral environment, when the
hypothalamic temperature equals T^ , both the panting and shivering motor neu
rons are more inhibited than facilitated, so there is no panting or shivering. As
the hypothalamic temperature increases, the facilitation of panting increases faster
than the inhibition. For temperatures above Tget , facilitation exceeds inhibition
and panting results in proportion to (Thypo — Tget )• Similarly, as the hypothalamic temperature drops below Th , inhibition of the motor neuron mediating
P&
** •Th may differ from the intrinsic hypothalamic set-point temperature Th^ ofFigure 1, since in
the wakeful dog in a neutral environment there may be and very likely are afferent inputs into the
hypothalamus from the thermal receptors in the skinand from the reticular activating system.
29
Cold Rteptort
Afftrtnt. from
Skin via Tholomut
Retplrolory Ana
(Panting)
Prtt.-Dtprof. Arto to SympiChain
(Cut. Vato-conttr.)
Sktlttal
Muscle
(Shivtring)
Firing
NEUTRAL
COLO
®I0
hi 0,0
HOT
lo Q
Sh\
3S
37
/pa
39
Sh\V-
35
41
10
\
37
39
41
39
37
39
41
Hypo. Tamp. 'C
Fig. 16—A physiological model for estoblishing a set-point temperature and illustrating possibili
ties for adjusting the set point. AC, anterior commissure; OC, optic chiasm; M, mammiilary body;
Pa, primary neuron for panting; Sh, primary neuron for shivering; v-c, primary neuron for vasoconstriction; crosshatched cell bodies, low-Q,o end high-Q„ primary sensory neurons. (Hammel,
H.T., 1965.)
shivering decreases more rapidly than does facilitation. For hypothalamic tem
peratures below Tget , facilitation exceeds inhibition, and shivering increases in
sh
proportion to (T^et - T^hypo)Sa
So far, we have considered a neuronal model of temperature regulation
located within the hypothalamus and have considered how it may fiinaion with
out input from outside itself, i.e., no input from peripheral receptors in the skin,
from extra hypothalamic core receptors, from the reticular activating system, or
from any other source. The model postulates that the primary sensory and pri-
30
mary neurons alone can activate thermoregulatory responses and, to do so, re
quire only an appropriate displacement of the hypothalamic temperature from
the fiinaional set-point temperature for each response. The model, thereby, con
forms with our experimental results obtained from displacing the hypothalamic
temperature and finding that the response is proportional to the difference be
tween the aaual hypothalamic temperature and the functional set-point tempera
ture for that response.
Our results then went on to show that placing the animal in hot or cold
environments did not actually lead to a useful displacement of the hypothalamic
temperature, although there were appropriate regulatory responses. We did not
choose to interpret this to mean that the hypothalamus is needlessly sensitive
to changes in its own temperature, but rather that an afferent input into the
hypothalamus from thermal receptors in the skin somehow shifts the functional
set-point temperature for each regulatory response. Our results obtained by dis
placing the hypothalamic temperature with the dog in neutral, cold, and hot
environments (Fig, 9), support this interpretation. They also suggest that the
environmental temperature shifts only the functional set-point temperatures and
not the proportionality constants.
The effects of afferent input into the hypothalamus consistent with our ex
perimental results may be readily and simply achieved by running afferent fibers
to the primary sensory neurons where they either facilitate or inhibit the spon
taneous and temperature-dependent activities of these sensory neurons. In Figure
16, afferents from the cold receptors in the skin are shown to facilitate the lo-Qio
sensory neurons and afferents from the warm receptors in the skin are shown to
facilitate the hi-Qio sensory neurons. In the cold environment, the activities of
the primary sensory and motor neurons are presumed to be as shown in the
lower left graph of Figure 16. The increased firing rate from the cold receptors
in the skin is shown to facilitate the activity of the lo-Qio sensory neurons, and
the decreased firing rate from the warm receptors in the skin is shown to reduce
the activity of the hi-Qio sensory neurons with the combined effect of raising
all functional set-point temperatures. Thus, although there may be no change
in the actual hypothalamic temperature—and, in fact, it may increase a little—it
is below the functional set-point temperatures for vasoconstriction and shivering
and will drive these responses in proportion to the differences.
In like manner, the activities of the primary sensory and motor neurons in
the hot environment are presumed to be as shown in the lower right graph of
Figure 16. Reduced firing rates from cold receptors in the skin are shown to re
duce the activity of the lo-Qio neurons, and increased firing rates from warm
receptors in the skin are shown to increase facilitation of the hi-Qio neurons,
with the combined effect of lowering all functional set-point temperatures. The
hypothalamic temperature, without changing, still drives panting and reduces
vasoconstriction.
31
Simply by suggesting that all afferent connections to the temperature regula
tory mechanism within the hypothalamus act by facilitating either the hi-Qio
or lo-Q,o sensory neurons, it is possible to account for the apparent shifts in the
functional set-point temperatures that occur in the transition from wakefulness
to sleep or in exercise, and without any apparent adjustment of the proportion
ality constants. For example, we may visualize that connections from the reticular activating system terminate on the lo-Qm sensory neurons and facilitate these
during the waking hours. At the onset of sleep, the facilitation may rapidly
diminish with a resultant immediate drop in all functional set-point tempera
tures—without changing any of the ttn's.
We recognize that terminating all inputs to the hypothalamus upon the
sensory neurons is not the only way to affect temperature regulation. It is pos
sible that some or all of the peripheral inputs feed into the motor neurons di
rectly and facilitate or inhibit these neurons. But to do so would require that
the peripheral inputs would also have to exercise antagonistic control over the
classes of primary motor neurons. For example, in order to shift all functional
set-point temperatures in the cold environment, as occurs experimentally, it is
necessary to suppose that the cold receptors in the skin not only facilitate the
primary neurons for shivering but at the same time inhibit the primary neurons
mediating panting. Similarly, the warm receptors in the skin must not only
facilitate the primary neurons for panting but also inhibit the neurons for shiver
ing. If the peripheral inputs do go directly to the primary motor neurons rather
than to the primary sensory neurons, then the hypothalamus may be thought of
as needlessly sensitive to changes in its own temperature because the hypothalamic temperature does not change in response to external thermal stress.
Since we know of no experimental evidence as to whether the afferent in
puts go directly to the primary motor neurons or go directly to the primary sen
sory neurons, we have favored the latter view because it is a simpler arrangement
and because it does not render the hypothalamus so needlessly sensitive to
changes in its own temperature.
It is worth noting that whatever afferent connections are made with the
hypothalamus, or for that matter whatever agent acts upon the primary sensory
or motor neurons within the hypothalamus, appears to affect regulation in the
same way—namely, by adjusting the functional set-point temperatures rather
than by changing the proportionality constants. Should a body of evidence ac
cumulate which demonstrates that the proportionality constants do change with
thermal stress, for example, Ofshivering increasing in cold exposure or oipanting
increasing in heat exposure, then it will be necessary to suggest that within a
class of primary motor neurons there is a wide range of thresholds or a range of
levels of spontaneous activity so that there is a range of functional set-point tem
peratures for each response. Thus, as the hypothalamic temperature deviates to
ward its extreme limits, there would be a recruitment of thermoregulatory re-
32
sponse and therefore increasing
At present, the evidence is too meager to
speculate further on this possibility.
I have discussed my opinion about the regulator of body temperature, and
formalized and expressed it again in terms of a speculative model of intercon
nected neurons. I would like to close with a discordant note. While heating the
hypothalamus of the dog in a cool environment, we always found, as noted in
Figure 10, that the falling rectal temperature was followed by an increase in
metabolism due to the onset of shivering. We interpreted this response to mean
that the lowered internal body temperature increased the threshold temperature
for shivering to a level above the temperature of the heated hypothalamus. We
should also note that the ear pinna vessels remain dilated even after shivering is
activated. Only after the thermal clamp was released did vasoconstriction occur
in the ear pinna. It is even possible to activate a more bizarre response. Extreme
cooling of the hypothalamus to 34® C in a cold environment will activate in
tense shivering and drive the internal body temperature rapidly up to 39.5° C.
If the environmental temperature is rapidly increased to 40® C while hypothalamic cooling continues, some dogs will repeatedly pant with interspersed bouts
of shivering at the same time. These unexpected results are lacking a satisfactory
explanation, but perhaps they are indicating a consequence of the technique we
have used for temperature displacement of the hypothalamus to which I have
already referred. When water of extreme temperatures are perfusing the thermode
surrounding the hypothalamus and when this acts to produce extreme internal
body (and blood) temperatures in the opposite direction, marked thermal gra
dients throughout the stimulated region may be expected. Because the pattern
of blood flow with respect to the thermodes may be highly irregular, we may
expect irregular gradients through the pre-optic tissue. Therefore it may not be
so surprising to find that contradictory regulatory responses are activated simul
taneously under such extreme thermal conditions. Obviously we should heat or
cool one side of the hypothalamus only or heat one side while cooling theother
in order to explore the consequences of temperature asymmetries within the
hypothalamus and its associated pre-optic region.
I am ending on this reference to an inherent fault in our technique, to in
dicate that I, too, am uncertain about our own effort to understand the regula
tor of body temperature.
33
Research in the author's laboratory is supported in part by
Contract AF-33-(657)-11103 with the 6570th Aerospace
Medical Research Laboratories, Wright-Patterson Air Force
Base, Ohio
34
REFERENCES
Andersson, B., C. C. Gale, and B. Hokfelt. 1964. Studies of the interaction between
neural and hormonal mechanisms in the regulation of body temperature, pp. 42-
61. In E. Bajusz and G. S.Jasmin (eds.) Major problems in neuroendocrinology.
The Williams and Wilkins Co., Baltimore.
Benzinger, T. H., C. Kitzinger, and A. "W. Pratt. 1963. The human thermostat, pp.
637-665. In Am. Inst. Physics. Temperature-its measurement and control in science
and industry. J. D. Hardy (ed.) Part 3, Biology and Medicine. Reinhold Pub
lishing Corp., New York.
Brody, S. 1945. Bioenergetics and growth. Reinhold Publ. Corp., New York.
Euler, C. von. 1961. Physiology and pharmacology of temperature regulation. Pharma
col. Rev. 13:361-398.
Hammel, H. T., C. H. Wyndham, and J. D. Hardy. 1958. Heat production and heat
loss in the dog at 8°-36° C environmental temperature. Am. J. Physiol., 194:99108.
Hammel, H. T., D. C. Jackson, J. A.J. Stolwijk, and J. D. Hardy. 1963. Hypothalamic temperatures in dog and monkey and thermoregulatory responses to environ
mental factors. USAF-TDR-63-5.
Hammel, H. T., D. C. Jackson, J. A. J. Stolwijk, J. D. Hardy, and S. B. Stromme.
1963. Temperature regulation by hypothalamic proportional control with an ad
justable set point. J. Appl. Physiol. 18:1146-1154.
Hammel, H. T., S. Stromme, and R. W. Cornew. 1963. Proportionality constant for
hypothalamic proportional control of metabolism in unanesthetized dog. Life Sci.
No. 2:933-947.
Hammel, H. T. 1965. Neurons and temperature regulation. Chapter 5 in Physiological
Controls and Regulations. Yamamoto and Brobeck (Ed.) Saunders, Philadelphia.
Hardy, J. D. 1961. Physiology of temperature regulation. Physiol. Rev. 41:521-606.
Hensel, H. and Zotterman, Y. 1951. Quantitative beziehungen zwischen der entladung
einzelner kaltefasern und der temperatur. Acta Physiol. Scand. 23:291.
Hensel, H., A. Iggo, and I. Witt. I960. A quantitative study of sensitive cutaneous
thermoreceptors with C afferent fibres. J. Physiol. 153:113-126.
Hensel, H. and K. K. A. Boman. I960. Afferent impulses in cutaneous sensory nerves
in human subjects. J. Neurophysiol. 23:564-578.
Jackson, D. C., and H. T. Hammel. 1963. Reduced set point temperature in exercising
dog. USAF AMRL-TDR-63-93.
Jackson, D. C. and H. T. Hammel. 1963. Hypothalamic set temperature decreased in
exercising dog. Life Sci. No. 8:554-563.
Jackson, D. C. 1963. Set point temperature as a factor in temperature regulation in
the exercising dog. Ph.D. Thesis. University of Pennsylvania.
Kuno, Y. 1956. Human perspiration. Charles C. Thomas, Publishers, Springfield, Illi
nois.
Nielsen, M. 1938. "Die Regulation der Korpertemperatur bei Muskelarbeit." Scand.
Arch. Physiol., 79, Suppl. 44:194.
Stolwijk, J. A.J. and J. D. Hardy. 1966. Partitional calorimetric studies of responses
of man to thermal transients. J. Appl. Physiol. 21:967-977.